Three groups of five rumen and duodenum cannulated fauna-free sheep were used in a 28 d experiment. One group remained fauna-free, whereas the second (EN) and third (PP) groups, respectively, were inoculated intraruminally with the protozoan species Entodinium caudatum and Polyplastron multivesiculatum. Rumen fluid, duodenal digesta and faecal samples were collected during the last 12 d. The flow of digesta to the duodenum was determined using Yb and Co as dual-phase markers. (15)Nitrogen and phosphatidylcholine were used as markers to calculate the duodenal flow of bacterial and protozoal N, respectively. Results showed an increase (P < 0.1) in the rumen concentration of NH3-N and total volatile fatty acids, and a decrease (P < 0.05) in the duodenal flow of non-NH3-N and bacterial N in sheep with EN and PP monofaunas, compared with fauna-free sheep. There were no differences (P > 0.05) in these variables between the two monofauna groups. Protozoal N accounted for 8% of the duodenal non-NH3-N flow in the EN-monofaunated sheep, whereas no such flow was detected in the PP-monofaunated sheep. Apparent rumen digestibility of organic matter, neutral detergent fibre and acid detergent fibre were similar (P > 0.05) in the monofaunated groups of sheep, but rumen acid detergent fibre digestibility was higher (P < 0.05) in the monofaunated than in the fauna-free groups. Experimental results suggested that, unlike EN, the PP monofauna might not contribute to the duodenal flow of microbial protein, whereas both monofaunas showed a virtually equal degree of predation on rumen bacteria.